The present experiments were de signed to investigate the role of housing and handling conditions during testing, as well as data analysis, on the outcome of antinociceptive testing of alpha(2)-adrenoceptor agonists, fentanyl, and a high dose of chlordiazepoxide in the tail withdrawal reaction test (TWR test) in rats. Dose-response curve data were obtained with fentanyl, clonidine, xylazine, dexmedetomidine, and 40.00 mg/kg chlordiazepoxide and were compared under normal TWR test conditions and during immobilization or immobilization with continuous painful stimulation. Data were analyzed in terms of ah-or-none criteria as well as percentage maximum possible effect (%MPE) analysis over the total measurement period or at any specific time point during testing. The results indicate that stress, induced by immobilization and immobilization with long-term-applied paw pressure, unmasked possible antinociceptive properties of the various alpha(2)-adrenoceptor agonists and potentiated the effects of fentanyl. Stress also unmasked the positive effects of benzodiazepines. The manner of data analysis was shown to significantly affect the outcome measured in stress and nonstress conditions. The MPE analysis, particularly at one time point, appeared much more sensitive than the ah-or-none criteria. The data indicate that the housing and handling conditions of animals during testing, together with data analysis, may affect the outcome of different classes of compounds in the TWR test, and this knowledge may help control for false positive results. (C) 1997 Elsevier Science Inc.